yade-dev team mailing list archive
-
yade-dev team
-
Mailing list archive
-
Message #03618
Re: [Branch ~yade-dev/yade/trunk] Rev 2079: Previous commit correction. Thanks to Vaclav.
Hi, Vaclav!
Thank you for fixing that. I have also found 2 days ago the difference
between Eigen and WM3 in my simulations, but, sorry, did not find the
reason.
I also need to solve the problem with clumps. Still have not found the
source of problems.
But I can only continue that work on April, when I return back to Germany.
Anyway, I also hope that Eigen migration will bring benefit :)
Thanks for help!
______________________________
Anton Gladkyy
2010/3/20 Václav Šmilauer <eudoxos@xxxxxxxx>
> > Hi Vaclav, thanks for suggestion, it seems, I have fixed it.
> Hi, what I suggested was wrong, as I saw from results of
> examples/concrete/uniax.py. I fixed that in the last commit. The results
> seem to be the same. There are still some things to wrap in python, run
> scripts/regression-tests.py, but I take care of that.
>
> The uniax.py simulation runs about 10x slower with eigen, probably due
> to lots of construction/descruction of temporary objects.
>
> Please test some of your simulations as well, then we will ask people to
> test theirs.
>
> Once this is done, we might either migrate yade code to eigen, or (which
> would be smoother and would let us compare performance), first
> progressively adapt miniWm3 interface to plain eigen (without our
> wrapper types) by adding extra methods, then changing sources of yade.
> We would create a few extra functions for things that have no direct
> correspondence in eigen, or ask upstream for including those.
>
> Long way ahead, but with solid basis of nice work of yours, it will be
> feasible, I think.
>
> Cheers, Vaclav
>
>
> >
> > When I ran some tests with old variant and a new one (with null-vector
> > checks), I got a little bit different results with regular-hexa. Does
> > it mean, that in some cases geometry was wrongly calculated
> > previously?
> >
> >
> > 2010/3/15 <noreply@xxxxxxxxxxxxx>
> > ------------------------------------------------------------
> > revno: 2079
> > committer: Anton Gladky <gladky.anton@xxxxxxxxx>
> > branch nick: trunk
> > timestamp: Mon 2010-03-15 12:24:35 +0100
> > message:
> > Previous commit correction. Thanks to Vaclav.
> > modified:
> >
> pkg/dem/DataClass/InteractionGeometry/Dem3DofGeom_SphereSphere.cpp
> >
> >
> > --
> > lp:yade
> > https://code.launchpad.net/~yade-dev/yade/trunk<https://code.launchpad.net/%7Eyade-dev/yade/trunk>
> >
> > Your team Yade developers is subscribed to branch lp:yade.
> > To unsubscribe from this branch go to
> >
> https://code.launchpad.net/~yade-dev/yade/trunk/+edit-subscription<https://code.launchpad.net/%7Eyade-dev/yade/trunk/+edit-subscription>
> .
> >
> > === modified file
> >
> 'pkg/dem/DataClass/InteractionGeometry/Dem3DofGeom_SphereSphere.cpp'
> > ---
> >
> pkg/dem/DataClass/InteractionGeometry/Dem3DofGeom_SphereSphere.cpp
> 2010-03-12 21:20:32 +0000
> > +++
> >
> pkg/dem/DataClass/InteractionGeometry/Dem3DofGeom_SphereSphere.cpp
> 2010-03-15 11:24:35 +0000
> > @@ -45,9 +45,10 @@
> > Vector3r axis=planeNormal.Cross(planePt);
> > axis.Normalize();
> > Real angle=planePt.Length()/radius;
> > normal2pt.FromAxisAngle(axis,angle);
> > + Quaternionr ret; return
> > ret.Align(Vector3r::UNIT_X,normal2pt*planeNormal);
> > + } else {
> > + return Quaternionr::IDENTITY;
> > }
> > - Quaternionr ret;
> > ret.Align(Vector3r::UNIT_X,normal2pt*planeNormal);
> > - return ret;
> > }
> >
> >
> >
> >
> > _______________________________________________
> > Mailing list: https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev>
> > Post to : yade-dev@xxxxxxxxxxxxxxxxxxx
> > Unsubscribe : https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev>
> > More help : https://help.launchpad.net/ListHelp
> >
> >
> > _______________________________________________
> > Mailing list: https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev>
> > Post to : yade-dev@xxxxxxxxxxxxxxxxxxx
> > Unsubscribe : https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev>
> > More help : https://help.launchpad.net/ListHelp
>
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev>
> Post to : yade-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev>
> More help : https://help.launchpad.net/ListHelp
>
References